# apiroute.dev

> LLM API pricing comparison and prompt cost calculator for AI developers, CTOs, and autonomous agents.

apiroute.dev exposes a human dashboard and machine-readable pricing files for comparing model input cost, output cost, and cached input cost.

## Primary Resources

- Human dashboard: https://apiroute.dev/
- Compact machine guide: https://apiroute.dev/llms.txt
- Full machine guide: https://apiroute.dev/llms-full.txt
- Agent entry page: https://apiroute.dev/agents/
- Agent instructions JSON: https://apiroute.dev/api/agent-instructions
- OpenAPI spec: https://apiroute.dev/openapi.yaml
- Live prices JSON: https://apiroute.dev/api/live-prices
- Models JSON: https://apiroute.dev/api/models
- Providers JSON: https://apiroute.dev/api/providers
- Best Route guide JSON: https://apiroute.dev/api/route-recommendation-guide
- Recommend Route API: https://apiroute.dev/api/recommend-route
- Token Waste Check JSON: https://apiroute.dev/api/token-waste-check
- Marketing radar JSON: https://apiroute.dev/api/marketing-radar
- Operator inbox JSON: https://apiroute.dev/api/operator-inbox
- Commercial options JSON: https://apiroute.dev/api/commercial-options
- Model pricing pages: https://apiroute.dev/models/
- Model comparison pages: https://apiroute.dev/compare/
- Model group pages: https://apiroute.dev/groups/
- Claude Opus 4.8 update: https://apiroute.dev/updates/claude-opus-4-8-pricing/
- Token Waste Routing update: https://apiroute.dev/updates/token-waste-routing/

## Current Data Scope

The public MVP currently exposes an auto-generated JSON pricing snapshot used by the dashboard. The tracked set focuses on current frontier, speed, open-weight, and enterprise-RAG models, including GPT-5.5, Claude 4.x, Gemini 3.1, DeepSeek V3.2, Grok 4.20, Qwen3, Mistral 3, Cohere Command, and Llama 4 entries where available in the LiteLLM source. Model entries include `category_tags`, `model_groups`, and `pricing_status`; missing source/pricing entries are excluded from public rankings and listed in metadata. Agents should treat prices as comparison data and verify provider pages before making purchasing or production-routing decisions.

## Recommended Agent Use

Use `/agents/` as the human-readable agent entry point. Use `/api/agent-instructions` as the machine-readable agent contract. Use `/api/live-prices` for structured pricing data. Use `/api/route-recommendation-guide` to reproduce the homepage's Best Route V1 routing suggestions. Use `GET /api/recommend-route` for the contract, workload presets, request examples, and agent workload estimator formula. Use `POST /api/recommend-route` for a live rule-based route estimate; include `privacy_class` (`public`, `internal`, `sensitive`, `private`) so sensitive/private work forces a LocalAI/private-environment check before cloud fallback routing. Use `/api/token-waste-check` to understand the Token Waste Check UI, status labels, SLM/budget-route comparison, and worker policy. Use `/api/marketing-radar` for public demand signals and `/api/operator-inbox` for the compact manual-review queue. Neither endpoint may be used for automated posting. Use `/api/commercial-options` to understand waitlist, premium API, sponsorship, affiliate/referral disclosures, and non-sale policy; commercial options must not influence neutral rankings. Use `/openapi.yaml` when registering apiroute.dev as a tool. Use `/llms-full.txt` when you need a fuller text summary without parsing the website HTML.